Forumite Members › General Topics › Tech › PC Talk › 5~30% CPU speed reduction to your Intel CPU forecast
- This topic has 44 replies, 10 voices, and was last updated 8 years ago by
Wheels-Of-Fire.
-
AuthorPosts
-
January 5, 2018 at 9:46 am #15281
Yes it was that number, but shows as apparently fully installed with no optional updates. My wife’s PC showed nothing due at all, though daughter’s machine does show updates pending. (The user is the challenge there, not the machine…)
All three run Home not pro versions which I understood allowed more leeway on updates.
January 5, 2018 at 10:44 am #15284Good job that you do not use Bitdefender then as apparently they have not yet updated their AV! link
edit – reading the small print unless a reg key is set the update will not be offered. So if you use MacCrappy or the like you will not get an update.
January 5, 2018 at 10:53 am #15286One of mine runs BitDefender Free A/V and it installed KB4056892 with no problems.
Never trust an atom - they make up everything !
January 5, 2018 at 11:32 am #15289One of mine runs BitDefender Free A/V and it installed KB4056892 with no problems.
That suggests that the update interferes with the A/V rather than the other way around, though other reports suggest that currently M$ are only slowing down server installations.
January 6, 2018 at 3:04 pm #15341El Reg reports that major CPU makers already have fixes for the ‘Meltdown’ back door. The Spectre flaw is currently unfixable at chip level.
The general details of how Spectre works are now available. Over-simplifying this is a side-channel attack that uses timing and training the CPU branch predictor to prefer one code branch over another then at a specific time slipping in a malign branch in its place. Apparently some Internet interfaces such as Chrome randomise timing results to make this harder – good work unfortunately completely undone by html5’s web-worker!
January 8, 2018 at 8:26 am #15459PcWorld have produced a very readable guide to Intel’s Meltdown Backdoor, while M$ have issued a PowerScript check for vulnerabilities. I have not linked the latter as unless you know what you are doing PowerScript can be a bit like handing a box of matches to a kiddy. In any case the M$ script is not much use, unless you have installed the Intel firmware update it will only show the Windows Slowdown status. Getting the firmware update is going to be a problem unless Intel break their usual policy and issue it via M$. The advice apparently is that your vendor should supply it – good luck with that!
I have bookmarked my link to the appropriate Intel i7 update release page for my own system and await this week’s developments.
[edit] Dave was correct my system has not slowed down but the power management on the chip shows a clock-rate speedup – this will cost me money!
January 11, 2018 at 10:07 am #15525Microsoft have detailed how their Meltdown fix impacts badly on older PCs. They state “most users will notice a decrease in system performance” for older machines. link
Although my 5+ years old i7 falls into that bracket, I did not notice a slowdown but I did notice a significant increase in power consumption as the cpu speed ramped up – I’m not a typical user in that I use VMs for most everything and they are well known i/o hogs sp perhaps my experience will not be reflected in that of others.
What I DO notice is that many Internet sites are a lot less responsive.
January 11, 2018 at 10:15 am #15526Ed, your link is no good, nowhere to go.
January 11, 2018 at 11:50 am #15530Seems that Nvidia chips are also now affected.
http://www.alphr.com/security/1008087/intel-chip-security-flaw-ARM-AMD
January 11, 2018 at 3:07 pm #15534Sorry about the link screw-up it should have pointed to this Verge article.
Intel Haswell ot older chips are the ones affected most by a slow-down.
January 12, 2018 at 2:17 pm #15568Oh dear Ed ?
More or less what I was afraid of.
Got an old i7 930 CUP
January 12, 2018 at 5:08 pm #15583Yep – looks like my workstation PC replacement will get accelerated a bit down to five years rather than seven, but luckily the vm bugs in AMD Ryzen now look like they are now sorted. ?
Probably mean me choosing a full AMD board and peripherals – the first time in over 20 years!
January 12, 2018 at 7:11 pm #15585Just fired up my X220 for the first time this week, I’m sure you can all guess what happened. ?
M$ are currently installing a massive bandwidth hog, my cable BB speed has dropped by 60%, Netflix had a hiccup with the quality and quickly sorted that out. If none of the family were on Windoze I’d nuke the X220 and put Linux, probably Mint, on.
January 12, 2018 at 8:33 pm #15588Unfortunately PM I think the Linux fixes will have a similar impact on old hardware. But if anyone knows better . . .
January 12, 2018 at 8:41 pm #15589The update is still going.
I know that the MS update counter is useless but it saying 36 hours to go.
Where’s that whisky I got for Christmas?
January 20, 2018 at 1:38 pm #15855Gibson Research has released a small tool called InSpectre that will tell you if your system firmware and OS have been patched.
Just Google Gibson Research and look in freeware.
January 20, 2018 at 2:03 pm #15856Many are still not fixed – Windows have only just released a Spectre fix for AMD.
That said, I am getting ready to migrate to AMD Ryzen from Intel so I’m doing a massive copying exercise — that is when you can easily see a 10-20% slowdown!
January 20, 2018 at 6:00 pm #15859My system tested as not patched on all counts ?.
Oh well. I will just have to hope that the anti virus people stay on top of any malware that may use these exploits. At least my browser says it’s patched ?
January 22, 2018 at 11:19 am #16013Linus Torvalds replies to a very critical message to Intel with an even more critical response! Most unusual for Linus to put in the boot to a chip company – such ire was only previously dished out only to the likes of Microsoft and Apple.
Pretty worrying for Intel investors as Intel seem to have completely lost the plot and are pouring petrol on the flames!
January 22, 2018 at 3:58 pm #16022If I read that correctly it seems Intel will be patching the problem in the microcode but not implementing it by default due to the performance hit?
So vendors will need to decide between security and ultimate performance in their products i.e. Intel’s doing just enough to avoid a class action.
If that is the case it could be AMD bonanza time. I wonder what ARM are doing?
-
AuthorPosts
- You must be logged in to reply to this topic.
